High Performance Ball Endmills - Inch Sizes - +/-.0002 Radii

[CBlair]

Well they do seem serious about documenting their work. Thanks for sharing, interesting information in this thread over some topics I would have never considered.

Charles

 

[CarbideBob]

Perhaps to clarify things.
Good starting rod comes in at +0/-.0003 inch.
Roundness is another tenth.
Every tool has some runout during the grind chucking so add another .00025 if your tooling in decent shape and replaced often or if checked before the grind cycle starts.
This runout must be checked at tool end and not at the collet so longer tools more a problem.

Add this up and your .5000 endmill will be made .4992 in diameter. But you will also have some mounting runout making it cut larger.
To not have the football shape which is a big no-no the tip rad needs to be .2496 or minus .0004 which is out of the ask.
In order to cut on the sides correctly the tool needs to be programmed at the .4992 diameter. Then the rad works out in 5 axis.
Obviously mixing tools made from ground rod and those made from oversize stock is a problem for comps.
In today's world the .002 error is huge but note that is .0008 at the max error point.
Done with machine vision (cameras) now by most. Unsure the new measuring devices past this 1980's stuff.
Why the two thou tolerance of old days?

We should welcome and endorse our new member with please stay with us and for sure shame me as needed.
Bob

 

[Jashley73]

Edit - Opps - I somehow realized that I'm late to the party ***again*** and that there were an extra 3-pages of replies between the first page, and this reply...


Carbidebob I think you and boosted may just be talking past each other.

What boosted is saying, is that there are far more CATALOG items available in metric sizes, for high-end, preimuim-tolerance tooling, than CATALOG available imperial sizes.

He's not saying that the machines magically become more capable of grinding an endmill to size & form in Dc 12.0000mm than they are capable of grinding to sise & form in Dc 12.7000mm. He's saying that those companies don't even bother to offer them in the Dc 12.7000 sizes.



Back on topic.

From memory, I remember that Mitsubishi makes some of their "Miracle" tools with +/-.002mm radius tolerance. These are tools that are sorted for tolerance, and the "fail" tools are sold as another (also very good) product series with a +/- .005mm tolerance. It would be worth a call to Mitsubishi to ask if these are offered in inch sizes.

I used to work for Seco, and remember that some of their Jabro ball-nose tools were being advertised as having +/- .005mm tolerances. Though I don't remember the specific series that was, Jabro does a fairly good job of supplying inch sizes, for the USA and UK markets. I would suggest a call to Seco Tools, and ask to speak to one of their solid-tools specialists. These guys have a much tighter relationship to the Niagara & Jabro factories, and would be the one's to talk to, in order to have strings pulled to get your tight-tolerance, sorted, ball-nose tools.

 

[CarbideBob]

Edit - Opps - I somehow realized that I'm late to the party ***again*** and that there were an extra 3-pages of replies between the first page, and this reply...
Carbidebob I think you and boosted may just be talking past each other.
.

Perhaps the thing to celebrate out of all this is that we now have sum-dum lured into the forum.
Obviously experienced.
Bob

 

[empower]

Perhaps the thing to celebrate out of all this is that we now have sum-dum lured into the forum.
Obviously experienced.
Bob

not my place to reveal his identity, but i can assure you he is an expert in this field.

 

[sum_dum_apps_guy]

Perhaps the thing to celebrate out of all this is that we now have sum-dum lured into the forum.
Obviously experienced.
Bob

One more place to spend my time avoiding the work I should be doing!
I also agree with your earlier statement that the ball nose in CAM should be programmed at 0.4992 (or whatever the manufacturer claims as the center of their tolerance).

 

[Shop Supply Guru]

DiJet makes indexable Inch and Metric Ball Nose & Bull Nose inserts. Here is a link Dijet Mirror Ball <-- Click.

They don't have their full finishing catalog on their website for some reason. PM me your contact and I can email it to you.

Claim .0004"+/- form accuracy when in the holder and .00024" on the insert itself. They do make tighter tolerance inserts with .0001" accuracy but they are all metric.

They have grades and geometries for steel, non ferrous, high temps and hardened.

Let me know if you need a distributor. It's small company good people. Products are made in Japan. They stock a lot of the ball nose items as it is their bread and butter.

 

[5 axis Fidia guy]

I will vouch for the above Dijet Mirror ball. It's THE best insertable ball endmill I have ever run, (I tested, OSG Pheonix, Fette, Dapra, Mits, Hitachi, Walther, and a few other oddballs). Dijets solid carbide ball endmills are a few notches above the rest as well. I still don't see where guys are finding all the endmill errors. I run the majority of my ball endmills thru the laser and actually check the form. Now granted the Hermle laser is not really made for a tool inspection device, it is however accurate within a micron or three. I would say MOST of the premium tooling I ran thru the laser are true to their word, within .0005 on the standards and .0002 on the "tight tolerance" tooling. I'm not saying the op is lying, I just can't imagine anyone grinding a tool .002 off.

 

[empower]

I will vouch for the above Dijet Mirror ball. It's THE best insertable ball endmill I have ever run, (I tested, OSG Pheonix, Fette, Dapra, Mits, Hitachi, Walther, and a few other oddballs). Dijets solid carbide ball endmills are a few notches above the rest as well. I still don't see where guys are finding all the endmill errors. I run the majority of my ball endmills thru the laser and actually check the form. Now granted the Hermle laser is not really made for a tool inspection device, it is however accurate within a micron or three. I would say MOST of the premium tooling I ran thru the laser are true to their word, within .0005 on the standards and .0002 on the "tight tolerance" tooling. I'm not saying the op is lying, I just can't imagine anyone grinding a tool .002 off.

how many points are you checking on the radius for form?

 

[5 axis Fidia guy]

I check 8 points around the radius, most of the tools I check are on the minus side. The Heidenhain control can actually use this chart and comp the 3d radius along the vectors to correct bogus shapes, even with a 5 axis toolpath.
It's actually crazy the control can figure this out, it took a long time to trust it but it indeed is pretty sweet, I am doing it as we speak, here is a little snippet of what the code looks like.



15179 LN X-0.30969 Y+2.68698 Z+2.18981 NX-0.991504545 NY+0.087686628 NZ+0.096071806 TX+0.015388753 TY+0.104375192 TZ+0.994418929
15180 LN X-0.31329 Y+2.687 Z+2.18983 NX-0.989248413 NY+0.100138768 NZ+0.106582382 TX+0.015265113 TY+0.104402123 TZ+0.994418007
15181 LN X-0.40097 Y+2.68749 Z+2.19047 NX-0.768963844 NY+0.479168193 NZ+0.423193158 TX+0.012268621 TY+0.105007737 TZ+0.994395724
15182 LN X-0.48894 Y+2.68799 Z+2.1911 NX-0.23813418 NY+0.738155916 NZ+0.631203578 TX+0.00925064 TY+0.105527274 TZ+0.994373381

 

[CBlair]

I check 8 points around the radius, most of the tools I check are on the minus side. The Heidenhain control can actually use this chart and comp the 3d radius along the vectors to correct bogus shapes, even with a 5 axis toolpath.

It is said that any technology suitably advanced is indistinguishable from magic...

As I work with a 20+ year old machine, and have worked with older ones it amazes me with what can be achieved. I didnt know that anyone did what you were talking about. On the machine in real time?

Charles

 

[sum_dum_apps_guy]

I check 8 points around the radius, most of the tools I check are on the minus side. The Heidenhain control can actually use this chart and comp the 3d radius along the vectors to correct bogus shapes, even with a 5 axis toolpath.
It's actually crazy the control can figure this out, it took a long time to trust it but it indeed is pretty sweet, I am doing it as we speak, here is a little snippet of what the code looks like.



15179 LN X-0.30969 Y+2.68698 Z+2.18981 NX-0.991504545 NY+0.087686628 NZ+0.096071806 TX+0.015388753 TY+0.104375192 TZ+0.994418929
15180 LN X-0.31329 Y+2.687 Z+2.18983 NX-0.989248413 NY+0.100138768 NZ+0.106582382 TX+0.015265113 TY+0.104402123 TZ+0.994418007
15181 LN X-0.40097 Y+2.68749 Z+2.19047 NX-0.768963844 NY+0.479168193 NZ+0.423193158 TX+0.012268621 TY+0.105007737 TZ+0.994395724
15182 LN X-0.48894 Y+2.68799 Z+2.1911 NX-0.23813418 NY+0.738155916 NZ+0.631203578 TX+0.00925064 TY+0.105527274 TZ+0.994373381

We have this on our Hermle as well. It is an impressive feature that I had to see to believe.

 

[5 axis Fidia guy]

It is said that any technology suitably advanced is indistinguishable from magic...

As I work with a 20+ year old machine, and have worked with older ones it amazes me with what can be achieved. I didnt know that anyone did what you were talking about. On the machine in real time?

Charles

Yes, real time, it's not like the feedrate slows down either, at least not that I can see.

 

[CarbideBob]

Yes, real time, it's not like the feedrate slows down either, at least not that I can see.

It's a measuring of the tool sort of like a cmm.
The offsets and rad now in and feed rate will not be affected.
May seem like magic but it is rather simple.
Compare the math used here to that in a good computer game.
One market is bigger.
Bob

 

[dstryr]

Must be the new Blum laser? My Grobs all have the NC4 which isn't capable of doing this...